Cancer and AI — Selected Events

How do LLMs work?

Mary had a little lamb 88.59%
Lamb 0.74%
lamp 0.65%
\n 0.40%
skirt 0.21%
Visualization adapted from VanderHart, “RDF and the future of LLMs” (ClojureTV, YouTube); top-5 next-token probabilities from OpenAI gpt-3.5-turbo-instruct, T=0
30 / 34